7 papers
Language Modeling is Monotone Compression
Noam Mazor, Andrew Morgan, Rafael Pass
A long-standing hypothesis in artificial intelligence and neuroscience posits that intelligence is closely related to compression: the ability to compress information efficiently i…
Can we Watermark Low-Entropy LLM Outputs?
Noam Mazor, Andrew Morgan, Rafael Pass
A recent and exciting thread of work focuses on developing methods for watermarking the output of large language models (LLMs). We focus on provably undetectable watermarking-that…
Computationally Differentially Private Inner Product Protocols Imply Oblivious Transfer
Iftach Haitner, Noam Mazor, Jad Silbak +2
In distributed differential privacy, multiple parties collaborate to analyze their combined data while each party protects the confidentiality of its data from the others. Interest…
On the Complexity of Two-Party Differential Privacy
Iftach Haitner, Noam Mazor, Jad Silbak +1
In distributed differential privacy, the parties perform analysis over their joint data while preserving the privacy for both datasets. Interestingly, for a few fundamental two-par…
Channels of Small Log-Ratio Leakage and Characterization of Two-Party Differentially Private Computation
Iftach Haitner, Noam Mazor, Ronen Shaltiel +1
Consider a PPT two-party protocol in which the parties get no private inputs and obtain outputs , and let and denote the parties' individu…
Lower Bounds on the Time/Memory Tradeoff of Function Inversion
Dror Chawin, Iftach Haitner, Noam Mazor
We study time/memory tradeoffs of function inversion: an algorithm, i.e., an inverter, equipped with an s-bit advice on a randomly chosen function and using or…
On the Communication Complexity of Key-Agreement Protocols
Iftach Haitner, Noam Mazor, Rotem Oshman +2
Key-agreement protocols whose security is proven in the random oracle model are an important alternative to protocols based on public-key cryptography. In the random oracle model,…